Initialize collective profiler with model config

This commit is contained in:
2026-07-16 21:59:58 +08:00
parent 3121e35b0e
commit 32ab584789
3 changed files with 23 additions and 5 deletions

View File

@@ -1,12 +1,12 @@
version = 1
[[jobs]]
name = "qwen30-vllm020-flashinfer-allreduce-smoke-tp2-20260716-v2-config-context"
name = "qwen30-vllm020-flashinfer-allreduce-smoke-tp2-20260716-v3-model-context"
gpus = 2
gpu_model = "H20"
hosts = ["dash0"]
command = "cd /home/admin/cpfs/wjh/aituner/aituner-qwen30-vllm020-profile-v1/runs/frontier-qwen30-vllm020-profile-v1 && timeout --signal=TERM --kill-after=30s 840 bash run_allreduce_profile.sh"
artifacts = ["artifacts/allreduce-smoke-tp2-v2"]
artifacts = ["artifacts/allreduce-smoke-tp2-v3"]
[jobs.env]
HOME = "/tmp/wjh"
@@ -14,6 +14,6 @@ XDG_CACHE_HOME = "/tmp/wjh/.cache"
VLLM_CACHE_ROOT = "/tmp/wjh/.cache/vllm"
TP = "2"
NUM_TOKENS = "8"
OUTPUT_ROOT = "/home/admin/cpfs/wjh/aituner/aituner-qwen30-vllm020-profile-fleet/artifacts/allreduce-smoke-tp2-v2"
OUTPUT_ROOT = "/home/admin/cpfs/wjh/aituner/aituner-qwen30-vllm020-profile-fleet/artifacts/allreduce-smoke-tp2-v3"
VENV_ROOT = "/tmp/wjh/venvs/vllm-0.20.0-cu129-profiler-v1"
VLLM_SOURCE = "/home/admin/cpfs/wjh/agentic-kv/third_party/vllm_v20_build"